scai 0.1.178 → 1.0.0

package/dist/testing/runtimeLogReader.js

@@ -0,0 +1,144 @@
+/**
+ * Reads runtime log text in both legacy banner format and newer shared-log format.
+ *
+ * Why this file exists:
+ * - evals and diagnosis still need to read old banner logs during migration
+ * - newer runtime entries no longer all use `INPUT | ...` / `OUTPUT | ...` blocks
+ * - one parser keeps report code from duplicating fragile log scraping rules
+ */
+function stripAnsi(text) {
+    return text.replace(/\x1B\[[0-9;]*m/g, "");
+}
+/**
+ * Normalizes runtime log text before eval token checks.
+ * Example: `[agent] [\x1b[33mtask:created\x1b[39m]` -> `[agent] [task:created]`.
+ */
+export function normalizeRuntimeLogText(text) {
+    return stripAnsi(text);
+}
+/**
+ * Matches both older ISO-UTC runtime headers and newer local-time human headers.
+ * Example: `[2026-04-30T12:00:00.000Z] ...` and `[2026-05-01 12:15:00 +02:00] ...`.
+ */
+function isStructuredRuntimeHeader(line) {
+    return /^\[[^\]]+\] \[[^\]]+\] \[[^:\]]+:[^\]]+\]/.test(normalizeRuntimeLogText(line));
+}
+function parseJsonIfPossible(text) {
+    const trimmed = text.trim();
+    if (!trimmed)
+        return "";
+    try {
+        return JSON.parse(trimmed);
+    }
+    catch {
+        return trimmed;
+    }
+}
+function splitStructuredRuntimeEntries(runLog) {
+    const lines = runLog.split("\n");
+    const entries = [];
+    let currentHeader;
+    let currentBody = [];
+    const flush = () => {
+        if (!currentHeader)
+            return;
+        entries.push({
+            header: currentHeader,
+            body: currentBody.join("\n").trim(),
+        });
+    };
+    for (const line of lines) {
+        if (isStructuredRuntimeHeader(line)) {
+            flush();
+            currentHeader = line;
+            currentBody = [];
+            continue;
+        }
+        if (currentHeader) {
+            currentBody.push(line);
+        }
+    }
+    flush();
+    return entries;
+}
+function extractLegacyBannerPayloads(runLog, stepName) {
+    const escapedName = stepName.replace(/[.*+?^${}()|[\]\\]/g, "\\$&");
+    const regex = new RegExp(`📂 OUTPUT \\| ${escapedName}[\\s\\S]*?\\n={68}\\n([\\s\\S]*?)\\n={68}\\n`, "g");
+    const payloads = [];
+    let match = null;
+    while ((match = regex.exec(runLog)) !== null) {
+        payloads.push(parseJsonIfPossible(match[1] ?? ""));
+    }
+    return payloads;
+}
+function extractStructuredLegacyPayloads(runLog, stepName) {
+    return splitStructuredRuntimeEntries(runLog)
+        .filter((entry) => entry.header.includes("[shared] [legacy:output]"))
+        .map((entry) => parseJsonIfPossible(entry.body))
+        .flatMap((payload) => {
+        if (typeof payload !== "object" || payload === null)
+            return [];
+        const legacy = payload;
+        if (legacy.payload?.type !== "output" || legacy.payload?.stepName !== stepName) {
+            return [];
+        }
+        return [legacy.payload.content];
+    });
+}
+export function extractStepPayloadsFromRuntimeLog(runLog, stepName) {
+    return [
+        ...extractLegacyBannerPayloads(runLog, stepName),
+        ...extractStructuredLegacyPayloads(runLog, stepName),
+    ];
+}
+function extractLegacyFinalAnswerText(runLog) {
+    const blockRegex = /📂 OUTPUT \| (?:finalAnswerModule|continuityAnswerModule|final-answer)[\s\S]*?\n={68}\n([\s\S]*?)\n={68}\n/g;
+    let lastBody = "";
+    let match = null;
+    while ((match = blockRegex.exec(runLog)) !== null) {
+        lastBody = match[1] ?? "";
+    }
+    if (!lastBody)
+        return "";
+    const parsed = parseJsonIfPossible(lastBody);
+    if (typeof parsed === "string")
+        return parsed;
+    if (typeof parsed === "object" && parsed !== null && "data" in parsed) {
+        const data = parsed.data;
+        return typeof data === "string" ? data : JSON.stringify(data ?? "");
+    }
+    return typeof parsed === "string" ? parsed : JSON.stringify(parsed);
+}
+export function extractFinalAnswerTextFromRuntimeLog(runLog) {
+    const legacyAnswer = extractLegacyFinalAnswerText(runLog);
+    if (legacyAnswer)
+        return legacyAnswer;
+    const structuredEntries = splitStructuredRuntimeEntries(runLog);
+    let latestAnswer = "";
+    for (const entry of structuredEntries) {
+        if (normalizeRuntimeLogText(entry.header).includes("artifact final-answer")) {
+            const parsed = parseJsonIfPossible(entry.body);
+            latestAnswer = typeof parsed === "string" ? parsed : JSON.stringify(parsed);
+            continue;
+        }
+        if (!normalizeRuntimeLogText(entry.header).includes("[shared] [legacy:output]"))
+            continue;
+        const parsed = parseJsonIfPossible(entry.body);
+        if (typeof parsed !== "object" || parsed === null)
+            continue;
+        const legacy = parsed;
+        if (legacy.payload?.stepName !== "finalAnswerModule" && legacy.payload?.stepName !== "continuityAnswerModule") {
+            continue;
+        }
+        const content = legacy.payload.content;
+        if (typeof content === "object" && content !== null && "data" in content) {
+            const data = content.data;
+            latestAnswer = typeof data === "string" ? data : JSON.stringify(data ?? "");
+        }
+        else if (typeof content === "string") {
+            latestAnswer = content;
+        }
+    }
+    return latestAnswer;
+}
+//# sourceMappingURL=runtimeLogReader.js.map

package/dist/testing/runtimeLogReader.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"runtimeLogReader.js","sourceRoot":"","sources":["../../src/testing/runtimeLogReader.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAOH,SAAS,SAAS,CAAC,IAAY;IAC7B,OAAO,IAAI,CAAC,OAAO,CAAC,iBAAiB,EAAE,EAAE,CAAC,CAAC;AAC7C,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,uBAAuB,CAAC,IAAY;IAClD,OAAO,SAAS,CAAC,IAAI,CAAC,CAAC;AACzB,CAAC;AAED;;;GAGG;AACH,SAAS,yBAAyB,CAAC,IAAY;IAC7C,OAAO,2CAA2C,CAAC,IAAI,CAAC,uBAAuB,CAAC,IAAI,CAAC,CAAC,CAAC;AACzF,CAAC;AAED,SAAS,mBAAmB,CAAC,IAAY;IACvC,MAAM,OAAO,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC;IAC5B,IAAI,CAAC,OAAO;QAAE,OAAO,EAAE,CAAC;IACxB,IAAI,CAAC;QACH,OAAO,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;IAC7B,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,OAAO,CAAC;IACjB,CAAC;AACH,CAAC;AAED,SAAS,6BAA6B,CAAC,MAAc;IACnD,MAAM,KAAK,GAAG,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IACjC,MAAM,OAAO,GAAsB,EAAE,CAAC;IACtC,IAAI,aAAiC,CAAC;IACtC,IAAI,WAAW,GAAa,EAAE,CAAC;IAE/B,MAAM,KAAK,GAAG,GAAG,EAAE;QACjB,IAAI,CAAC,aAAa;YAAE,OAAO;QAC3B,OAAO,CAAC,IAAI,CAAC;YACX,MAAM,EAAE,aAAa;YACrB,IAAI,EAAE,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE;SACpC,CAAC,CAAC;IACL,CAAC,CAAC;IAEF,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,IAAI,yBAAyB,CAAC,IAAI,CAAC,EAAE,CAAC;YACpC,KAAK,EAAE,CAAC;YACR,aAAa,GAAG,IAAI,CAAC;YACrB,WAAW,GAAG,EAAE,CAAC;YACjB,SAAS;QACX,CAAC;QACD,IAAI,aAAa,EAAE,CAAC;YAClB,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QACzB,CAAC;IACH,CAAC;IAED,KAAK,EAAE,CAAC;IACR,OAAO,OAAO,CAAC;AACjB,CAAC;AAED,SAAS,2BAA2B,CAAC,MAAc,EAAE,QAAgB;IACnE,MAAM,WAAW,GAAG,QAAQ,CAAC,OAAO,CAAC,qBAAqB,EAAE,MAAM,CAAC,CAAC;IACpE,MAAM,KAAK,GAAG,IAAI,MAAM,CACtB,iBAAiB,WAAW,8CAA8C,EAC1E,GAAG,CACJ,CAAC;IACF,MAAM,QAAQ,GAAc,EAAE,CAAC;IAC/B,IAAI,KAAK,GAA2B,IAAI,CAAC;IACzC,OAAO,CAAC,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;QAC7C,QAAQ,CAAC,IAAI,CAAC,mBAAmB,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC;IACrD,CAAC;IACD,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED,SAAS,+BAA+B,CAAC,MAAc,EAAE,QAAgB;IACvE,OAAO,6BAA6B,CAAC,MAAM,CAAC;SACzC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,MAAM,CAAC,QAAQ,CAAC,0BAA0B,CAAC,CAAC;SACpE,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,mBAAmB,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;SAC/C,OAAO,CAAC,CAAC,OAAO,EAAE,EAAE;QACnB,IAAI,OAAO,OAAO,KAAK,QAAQ,IAAI,OAAO,KAAK,IAAI;YAAE,OAAO,EAAE,CAAC;QAC/D,MAAM,MAAM,GAAG,OAMd,CAAC;QACF,IAAI,MAAM,CAAC,OAAO,EAAE,IAAI,KAAK,QAAQ,IAAI,MAAM,CAAC,OAAO,EAAE,QAAQ,KAAK,QAAQ,EAAE,CAAC;YAC/E,OAAO,EAAE,CAAC;QACZ,CAAC;QACD,OAAO,CAAC,MAAM,CAAC,OAAO,CAAC,OAAO,CAAC,CAAC;IAClC,CAAC,CAAC,CAAC;AACP,CAAC;AAED,MAAM,UAAU,iCAAiC,CAAC,MAAc,EAAE,QAAgB;IAChF,OAAO;QACL,GAAG,2BAA2B,CAAC,MAAM,EAAE,QAAQ,CAAC;QAChD,GAAG,+BAA+B,CAAC,MAAM,EAAE,QAAQ,CAAC;KACrD,CAAC;AACJ,CAAC;AAED,SAAS,4BAA4B,CAAC,MAAc;IAClD,MAAM,UAAU,GACd,6GAA6G,CAAC;IAChH,IAAI,QAAQ,GAAG,EAAE,CAAC;IAClB,IAAI,KAAK,GAA2B,IAAI,CAAC;IACzC,OAAO,CAAC,KAAK,GAAG,UAAU,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;QAClD,QAAQ,GAAG,KAAK,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;IAC5B,CAAC;IACD,IAAI,CAAC,QAAQ;QAAE,OAAO,EAAE,CAAC;IACzB,MAAM,MAAM,GAAG,mBAAmB,CAAC,QAAQ,CAAC,CAAC;IAC7C,IAAI,OAAO,MAAM,KAAK,QAAQ;QAAE,OAAO,MAAM,CAAC;IAC9C,IAAI,OAAO,MAAM,KAAK,QAAQ,IAAI,MAAM,KAAK,IAAI,IAAI,MAAM,IAAI,MAAM,EAAE,CAAC;QACtE,MAAM,IAAI,GAAI,MAA6B,CAAC,IAAI,CAAC;QACjD,OAAO,OAAO,IAAI,KAAK,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,IAAI,EAAE,CAAC,CAAC;IACtE,CAAC;IACD,OAAO,OAAO,MAAM,KAAK,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,IAAI,CAAC,SAAS,CAAC,MAAM,CAAC,CAAC;AACtE,CAAC;AAED,MAAM,UAAU,oCAAoC,CAAC,MAAc;IACjE,MAAM,YAAY,GAAG,4BAA4B,CAAC,MAAM,CAAC,CAAC;IAC1D,IAAI,YAAY;QAAE,OAAO,YAAY,CAAC;IAEtC,MAAM,iBAAiB,GAAG,6BAA6B,CAAC,MAAM,CAAC,CAAC;IAChE,IAAI,YAAY,GAAG,EAAE,CAAC;IACtB,KAAK,MAAM,KAAK,IAAI,iBAAiB,EAAE,CAAC;QACtC,IAAI,uBAAuB,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,QAAQ,CAAC,uBAAuB,CAAC,EAAE,CAAC;YAC5E,MAAM,MAAM,GAAG,mBAAmB,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;YAC/C,YAAY,GAAG,OAAO,MAAM,KAAK,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,IAAI,CAAC,SAAS,CAAC,MAAM,CAAC,CAAC;YAC5E,SAAS;QACX,CAAC;QACD,IAAI,CAAC,uBAAuB,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,QAAQ,CAAC,0BAA0B,CAAC;YAAE,SAAS;QAC1F,MAAM,MAAM,GAAG,mBAAmB,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QAC/C,IAAI,OAAO,MAAM,KAAK,QAAQ,IAAI,MAAM,KAAK,IAAI;YAAE,SAAS;QAC5D,MAAM,MAAM,GAAG,MAKd,CAAC;QACF,IAAI,MAAM,CAAC,OAAO,EAAE,QAAQ,KAAK,mBAAmB,IAAI,MAAM,CAAC,OAAO,EAAE,QAAQ,KAAK,wBAAwB,EAAE,CAAC;YAC9G,SAAS;QACX,CAAC;QACD,MAAM,OAAO,GAAG,MAAM,CAAC,OAAO,CAAC,OAAO,CAAC;QACvC,IAAI,OAAO,OAAO,KAAK,QAAQ,IAAI,OAAO,KAAK,IAAI,IAAI,MAAM,IAAI,OAAO,EAAE,CAAC;YACzE,MAAM,IAAI,GAAI,OAA8B,CAAC,IAAI,CAAC;YAClD,YAAY,GAAG,OAAO,IAAI,KAAK,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,IAAI,EAAE,CAAC,CAAC;QAC9E,CAAC;aAAM,IAAI,OAAO,OAAO,KAAK,QAAQ,EAAE,CAAC;YACvC,YAAY,GAAG,OAAO,CAAC;QACzB,CAAC;IACH,CAAC;IACD,OAAO,YAAY,CAAC;AACtB,CAAC","sourcesContent":["/**\n * Reads runtime log text in both legacy banner format and newer shared-log format.\n *\n * Why this file exists:\n * - evals and diagnosis still need to read old banner logs during migration\n * - newer runtime entries no longer all use `INPUT | ...` / `OUTPUT | ...` blocks\n * - one parser keeps report code from duplicating fragile log scraping rules\n */\n\ntype RuntimeLogEntry = {\n  header: string;\n  body: string;\n};\n\nfunction stripAnsi(text: string): string {\n  return text.replace(/\\x1B\\[[0-9;]*m/g, \"\");\n}\n\n/**\n * Normalizes runtime log text before eval token checks.\n * Example: `[agent] [\\x1b[33mtask:created\\x1b[39m]` -> `[agent] [task:created]`.\n */\nexport function normalizeRuntimeLogText(text: string): string {\n  return stripAnsi(text);\n}\n\n/**\n * Matches both older ISO-UTC runtime headers and newer local-time human headers.\n * Example: `[2026-04-30T12:00:00.000Z] ...` and `[2026-05-01 12:15:00 +02:00] ...`.\n */\nfunction isStructuredRuntimeHeader(line: string): boolean {\n  return /^\\[[^\\]]+\\] \\[[^\\]]+\\] \\[[^:\\]]+:[^\\]]+\\]/.test(normalizeRuntimeLogText(line));\n}\n\nfunction parseJsonIfPossible(text: string): unknown {\n  const trimmed = text.trim();\n  if (!trimmed) return \"\";\n  try {\n    return JSON.parse(trimmed);\n  } catch {\n    return trimmed;\n  }\n}\n\nfunction splitStructuredRuntimeEntries(runLog: string): RuntimeLogEntry[] {\n  const lines = runLog.split(\"\\n\");\n  const entries: RuntimeLogEntry[] = [];\n  let currentHeader: string | undefined;\n  let currentBody: string[] = [];\n\n  const flush = () => {\n    if (!currentHeader) return;\n    entries.push({\n      header: currentHeader,\n      body: currentBody.join(\"\\n\").trim(),\n    });\n  };\n\n  for (const line of lines) {\n    if (isStructuredRuntimeHeader(line)) {\n      flush();\n      currentHeader = line;\n      currentBody = [];\n      continue;\n    }\n    if (currentHeader) {\n      currentBody.push(line);\n    }\n  }\n\n  flush();\n  return entries;\n}\n\nfunction extractLegacyBannerPayloads(runLog: string, stepName: string): unknown[] {\n  const escapedName = stepName.replace(/[.*+?^${}()|[\\]\\\\]/g, \"\\\\$&\");\n  const regex = new RegExp(\n    `📂 OUTPUT \\\\| ${escapedName}[\\\\s\\\\S]*?\\\\n={68}\\\\n([\\\\s\\\\S]*?)\\\\n={68}\\\\n`,\n    \"g\"\n  );\n  const payloads: unknown[] = [];\n  let match: RegExpExecArray | null = null;\n  while ((match = regex.exec(runLog)) !== null) {\n    payloads.push(parseJsonIfPossible(match[1] ?? \"\"));\n  }\n  return payloads;\n}\n\nfunction extractStructuredLegacyPayloads(runLog: string, stepName: string): unknown[] {\n  return splitStructuredRuntimeEntries(runLog)\n    .filter((entry) => entry.header.includes(\"[shared] [legacy:output]\"))\n    .map((entry) => parseJsonIfPossible(entry.body))\n    .flatMap((payload) => {\n      if (typeof payload !== \"object\" || payload === null) return [];\n      const legacy = payload as {\n        payload?: {\n          stepName?: unknown;\n          type?: unknown;\n          content?: unknown;\n        };\n      };\n      if (legacy.payload?.type !== \"output\" || legacy.payload?.stepName !== stepName) {\n        return [];\n      }\n      return [legacy.payload.content];\n    });\n}\n\nexport function extractStepPayloadsFromRuntimeLog(runLog: string, stepName: string): unknown[] {\n  return [\n    ...extractLegacyBannerPayloads(runLog, stepName),\n    ...extractStructuredLegacyPayloads(runLog, stepName),\n  ];\n}\n\nfunction extractLegacyFinalAnswerText(runLog: string): string {\n  const blockRegex =\n    /📂 OUTPUT \\| (?:finalAnswerModule|continuityAnswerModule|final-answer)[\\s\\S]*?\\n={68}\\n([\\s\\S]*?)\\n={68}\\n/g;\n  let lastBody = \"\";\n  let match: RegExpExecArray | null = null;\n  while ((match = blockRegex.exec(runLog)) !== null) {\n    lastBody = match[1] ?? \"\";\n  }\n  if (!lastBody) return \"\";\n  const parsed = parseJsonIfPossible(lastBody);\n  if (typeof parsed === \"string\") return parsed;\n  if (typeof parsed === \"object\" && parsed !== null && \"data\" in parsed) {\n    const data = (parsed as { data?: unknown }).data;\n    return typeof data === \"string\" ? data : JSON.stringify(data ?? \"\");\n  }\n  return typeof parsed === \"string\" ? parsed : JSON.stringify(parsed);\n}\n\nexport function extractFinalAnswerTextFromRuntimeLog(runLog: string): string {\n  const legacyAnswer = extractLegacyFinalAnswerText(runLog);\n  if (legacyAnswer) return legacyAnswer;\n\n  const structuredEntries = splitStructuredRuntimeEntries(runLog);\n  let latestAnswer = \"\";\n  for (const entry of structuredEntries) {\n    if (normalizeRuntimeLogText(entry.header).includes(\"artifact final-answer\")) {\n      const parsed = parseJsonIfPossible(entry.body);\n      latestAnswer = typeof parsed === \"string\" ? parsed : JSON.stringify(parsed);\n      continue;\n    }\n    if (!normalizeRuntimeLogText(entry.header).includes(\"[shared] [legacy:output]\")) continue;\n    const parsed = parseJsonIfPossible(entry.body);\n    if (typeof parsed !== \"object\" || parsed === null) continue;\n    const legacy = parsed as {\n      payload?: {\n        stepName?: unknown;\n        content?: unknown;\n      };\n    };\n    if (legacy.payload?.stepName !== \"finalAnswerModule\" && legacy.payload?.stepName !== \"continuityAnswerModule\") {\n      continue;\n    }\n    const content = legacy.payload.content;\n    if (typeof content === \"object\" && content !== null && \"data\" in content) {\n      const data = (content as { data?: unknown }).data;\n      latestAnswer = typeof data === \"string\" ? data : JSON.stringify(data ?? \"\");\n    } else if (typeof content === \"string\") {\n      latestAnswer = content;\n    }\n  }\n  return latestAnswer;\n}\n"]}

package/dist/testing/testCommands.js

@@ -1,310 +1,42 @@
-import fs from "fs";
+/**
+ * Registers dev-only shell test helpers.
+ *
+ * Why this file exists:
+ * - dev shell commands should stay in one place instead of being spread across the shell setup
+ * - test commands should stay in one place instead of being spread across shell setup
+ * - the shell should only show explicit maintainer helpers when dev mode is enabled
+ */
 import chalk from "chalk";
-import { RUN_LOG_PATH } from "../constants.js";
-// ---------------- Test Queries ----------------
-export const testQueries = [
-    "Add concise comments to semanticAnalysisModule.ts and finalAnswerModule.ts describing phase boundaries.",
-    "Refactor MainAgent runVerify flow to reduce nesting while preserving behavior.",
-    "Explain how resolveExecutionModeStep, routingDecisionStep, and canExecuteRoute interact.",
-    "Add stronger validation and safer fallback behavior in contextReviewStep.ts.",
-    "Summarize CLI architecture from index.ts, commands/factory.ts, and runQueryWithDaemonControl.ts.",
-    "Where are SQLite queries defined in db/fileIndex.ts and db/client.ts?",
-    "Which agent step files under cli/src/agents are missing test coverage?",
-    "What parts of runWorkLoop could cause repeated pending task steps?",
-    "How does runSearch seed relatedFiles and how does verify prune missing paths?",
-    "Refactor buildContextualPrompt.ts into smaller focused helpers without behavior change.",
-    "Is error handling consistent across MainAgent.ts, evidenceVerifierStep.ts, and semanticAnalysisModule.ts?",
-    "What important usage examples are missing from README for shell commands and /edit flow?",
-    "How does this repo run tests and where are the test entry points configured?",
-    "Are there flaky-test signals in __tests__ or test scripts, and where would retries be added?",
-    "Is there dead code in cli/src/agents that is never referenced by MainAgent or command flow?",
-    "Map the full run lifecycle in MainAgent and point out where routing decisions are made.",
-    "For repo-wide analysis questions, which method decides verify wave budget and why?",
-    "Compare runSearch, runVerify, and runResearch responsibilities and identify overlap risks.",
-    "Trace how selectedFiles and candidateFiles evolve across verify waves for this query.",
-    "Show where taskStep status can get stuck pending and propose a bounded completion rule.",
-    "Refactor evidenceVerifierStep so verify data is separated from semantic analysis outputs.",
-    "Identify all places where action.shouldModify is set and flag analyze-mode violations.",
-    "Which modules still read FileAnalysis.evidence or searchScore after verify-state split?",
-    "Given a multi-question prompt, how should planning split research vs execution steps?",
-    "Find the minimal file set required to answer: how do I run tests and where are flaky tests detected?",
-    "Explain why a repo-wide question might still end up analyzing only a few files.",
-    "List all gate checks (phase/scope/route/readiness/research) and the exact stop conditions.",
-    "Audit finalAnswer analyzed-files output for mismatch against actually executed taskSteps.",
-    "Where does semanticAnalysis merge prior state, and what fields should remain semantic-only?",
-];
-// ---------------- Local Evals ----------------
-export const localEvalCases = [
-    // ---------------- Routing / Verify ----------------
-    {
-        id: "routing_repo_wide_question",
-        query: "summarize this repo architecture and identify weak coupling points",
-        checks: {
-            mustContain: [
-                "OUTPUT | readinessGateStep",
-                "OUTPUT | runResearch",
-                "OUTPUT | reasonNextTaskStep",
-            ],
-        },
-    },
-    {
-        id: "verify_pipeline_sqlite_query",
-        query: "Where are SQLite queries defined in db/fileIndex.ts and db/client.ts?",
-        checks: {
-            mustContain: [
-                "OUTPUT | evidenceVerifier",
-                "OUTPUT | preFileSearchCheckStep",
-                "OUTPUT | readinessGateStep",
-            ],
-        },
-    },
-    {
-        id: "verify_focus_evolution",
-        query: "Trace how selectedFiles and candidateFiles evolve across verify waves for this query.",
-        checks: {
-            mustContain: [
-                "OUTPUT | evidenceVerifier",
-                "OUTPUT | preFileSearchCheckStep",
-                "OUTPUT | selectRelevantSources",
-            ],
-        },
-    },
-    // ---------------- Research ----------------
-    {
-        id: "research_generation_present",
-        query: "Is error handling consistent across the codebase and how do I run the test suite?",
-        when: {
-            mustContainInRunLog: ['"allowResearch": true'],
-        },
-        checks: {
-            mustContain: [
-                "OUTPUT | researchPlanGen",
-                "research-impact-map",
-                "research-symbol-trace",
-                "research-risk-check",
-            ],
-        },
-    },
-    {
-        id: "research_execution_artifacts_present",
-        query: "Is error handling consistent across the codebase and how do I run the test suite?",
-        when: {
-            mustContainInRunLog: ['"allowResearch": true'],
-        },
-        checks: {
-            mustContain: [
-                "OUTPUT | runResearch",
-                "OUTPUT | runResearchStep",
-                "\"collectedData\"",
-            ],
-        },
-    },
-    {
-        id: "research_synthesis_present",
-        query: "summarize this repo architecture and identify weak coupling points",
-        when: {
-            mustContainInRunLog: ['"allowResearch": true'],
-        },
-        checks: {
-            mustContain: [
-                "research-architecture-synthesis",
-                "OUTPUT | runResearchStep",
-                "\"problemStatement\"",
-            ],
-        },
-    },
-    // ---------------- Workloop / Analysis ----------------
-    {
-        id: "verify_store_used",
-        query: "For repo-wide analysis questions, which method decides verify wave budget and why?",
-        checks: {
-            mustContain: [
-                "OUTPUT | evidenceVerifier",
-                "OUTPUT | preFileSearchCheckStep",
-            ],
-        },
-    },
-    {
-        id: "analysis_plan_and_semantic_pass",
-        query: "Where does semanticAnalysis merge prior state, and what fields should remain semantic-only?",
-        checks: {
-            mustContain: [
-                "OUTPUT | analysisPlanGen",
-                "OUTPUT | semanticAnalysisStep - per-file",
-                "OUTPUT | reasonNextStep",
-            ],
-        },
-    },
-    {
-        id: "loop_safety_for_pending_steps",
-        query: "Show where taskStep status can get stuck pending and propose a bounded completion rule.",
-        checks: {
-            mustContain: [
-                "OUTPUT | reasonNextTaskStep",
-                "OUTPUT | reasonNextStep",
-            ],
-            maxOccurrences: {
-                "OUTPUT | reasonNextStep": 40,
-            },
-        },
-    },
-    // ---------------- Final Answer ----------------
-    {
-        id: "final_answer_for_test_entrypoints",
-        query: "How does this repo run tests and where are the test entry points configured?",
-        checks: {
-            mustContain: [
-                "OUTPUT | finalAnswerModule",
-            ],
-            mustContainAnswer: [
-                "test",
-            ],
-        },
-    },
-    {
-        id: "final_answer_analyzed_files_reported",
-        query: "Audit finalAnswer analyzed-files output for mismatch against actually executed taskSteps.",
-        checks: {
-            mustContain: [
-                "OUTPUT | finalAnswerModule",
-                "\"analyzedFiles\"",
-            ],
-        },
-    },
-];
-function pickRandom(items) {
-    return items[Math.floor(Math.random() * items.length)];
-}
-function logHighContrastQuery(prefix, query) {
-    // Neutral bold so it is readable on both light and dark terminal themes.
-    console.log(chalk.bold(`${prefix} ${query}`));
-}
-function countOccurrences(haystack, needle) {
-    if (!needle)
-        return 0;
-    return haystack.split(needle).length - 1;
-}
-function extractFinalAnswerText(runLog) {
-    const blockRegex = /📂 OUTPUT \| finalAnswerModule[\s\S]*?\n={68}\n([\s\S]*?)\n={68}\n/g;
-    let lastBody = "";
-    let match = null;
-    while ((match = blockRegex.exec(runLog)) !== null) {
-        lastBody = match[1] ?? "";
-    }
-    if (!lastBody)
-        return "";
-    try {
-        const parsed = JSON.parse(lastBody.trim());
-        if (typeof parsed?.data === "string")
-            return parsed.data;
-        return JSON.stringify(parsed?.data ?? "");
-    }
-    catch {
-        return lastBody;
-    }
-}
-function isEvalApplicable(test, runLog) {
-    const mustContain = test.when?.mustContainInRunLog ?? [];
-    for (const token of mustContain) {
-        if (!runLog.includes(token)) {
-            return { ok: false, reason: `condition unmet: missing "${token}"` };
-        }
-    }
-    const mustNotContain = test.when?.mustNotContainInRunLog ?? [];
-    for (const token of mustNotContain) {
-        if (runLog.includes(token)) {
-            return { ok: false, reason: `condition unmet: found forbidden "${token}"` };
-        }
-    }
-    return { ok: true };
-}
-async function runLocalEvals(runQuery) {
-    const startedAt = Date.now();
-    let passCount = 0;
-    let skipCount = 0;
-    for (const test of localEvalCases) {
-        fs.mkdirSync(requireRunLogDir(), { recursive: true });
-        fs.writeFileSync(RUN_LOG_PATH, "", { flag: "w" });
-        console.log(chalk.cyan(`\n[EVAL] ${test.id}`));
-        logHighContrastQuery("[EVAL QUERY]", test.query);
-        fs.appendFileSync(RUN_LOG_PATH, `\n[EVAL] ${test.id}\n[EVAL QUERY] ${test.query}\n`, "utf-8");
-        await runQuery(test.query);
-        const runLog = fs.existsSync(RUN_LOG_PATH)
-            ? fs.readFileSync(RUN_LOG_PATH, "utf-8")
-            : "";
-        const finalAnswerText = extractFinalAnswerText(runLog);
-        const applicability = isEvalApplicable(test, runLog);
-        if (!applicability.ok) {
-            skipCount++;
-            console.log(chalk.blue(`[SKIP] ${test.id} (${applicability.reason})`));
-            continue;
-        }
-        const failures = [];
-        for (const token of test.checks.mustContain ?? []) {
-            if (!runLog.includes(token)) {
-                failures.push(`missing token: ${token}`);
-            }
-        }
-        for (const token of test.checks.mustNotContain ?? []) {
-            if (runLog.includes(token)) {
-                failures.push(`unexpected token: ${token}`);
-            }
-        }
-        for (const [token, max] of Object.entries(test.checks.maxOccurrences ?? {})) {
-            const count = countOccurrences(runLog, token);
-            if (count > max) {
-                failures.push(`too many occurrences: "${token}" => ${count} (max ${max})`);
-            }
-        }
-        for (const token of test.checks.mustContainAnswer ?? []) {
-            if (!finalAnswerText.toLowerCase().includes(token.toLowerCase())) {
-                failures.push(`final answer missing token: ${token}`);
-            }
-        }
-        for (const token of test.checks.mustNotContainAnswer ?? []) {
-            if (finalAnswerText.toLowerCase().includes(token.toLowerCase())) {
-                failures.push(`final answer contains forbidden token: ${token}`);
-            }
-        }
-        if (failures.length === 0) {
-            passCount++;
-            console.log(chalk.green(`[PASS] ${test.id}`));
-        }
-        else {
-            console.log(chalk.red(`[FAIL] ${test.id}`));
-            for (const failure of failures) {
-                console.log(chalk.red(`  - ${failure}`));
-            }
-        }
-    }
-    const durationSec = ((Date.now() - startedAt) / 1000).toFixed(1);
-    const total = localEvalCases.length;
-    const failed = total - passCount - skipCount;
-    const color = failed === 0 ? chalk.green : chalk.yellow;
-    console.log(color(`\n[EVAL SUMMARY] pass=${passCount}/${total}, skip=${skipCount}, fail=${failed}, duration=${durationSec}s`));
-}
-function requireRunLogDir() {
-    const idx = RUN_LOG_PATH.lastIndexOf("/");
-    if (idx <= 0)
-        return ".";
-    return RUN_LOG_PATH.slice(0, idx);
-}
+import { runAllTestsCommand, listRegisteredTests, printTestCommand, runRegisteredTestCommand, } from "../commands/TestCmd.js";
 /**
- * Registers testing-related shell commands.
- * Example: /test, /test-random, /test-evals.
+ * Registers shell test commands.
+ * Example: /tests, /test-run <id>.
+ *
+ * To run tests:
+ * - Use `/tests` to run the full test suite.
+ * - Use `/test-run <id>` to run a specific test by its ID.
+ * - Use `/test-list` to list all available tests.
+ * - Use `/test-cmd [id]` to print the command for the full suite or a specific test.
  */
-export function registerTestingCommands(customCommands, runQuery) {
-    customCommands.test = async () => {
-        const query = testQueries[0];
-        logHighContrastQuery("[TEST QUERY]", query);
-        await runQuery(query);
+export function registerTestingCommands(customCommands, _runQuery) {
+    customCommands.tests = async () => {
+        runAllTestsCommand();
+    };
+    customCommands["test-run"] = async (...args) => {
+        const testId = (args[0] ?? "").trim();
+        if (!testId) {
+            console.log(chalk.yellow("Usage: /test-run <id>"));
+            listRegisteredTests();
+            return;
+        }
+        runRegisteredTestCommand(testId);
     };
-    customCommands["test-random"] = async () => {
-        const query = pickRandom(testQueries);
-        console.log(`\n🎲 [test-random] Selected query:\n→ ${query}\n`);
-        await runQuery(query);
+    customCommands["test-list"] = async () => {
+        listRegisteredTests();
     };
-    customCommands["test-evals"] = async () => {
-        await runLocalEvals(runQuery);
+    customCommands["test-cmd"] = async (...args) => {
+        const testId = (args[0] ?? "").trim();
+        printTestCommand(testId || undefined);
     };
 }
+//# sourceMappingURL=testCommands.js.map

package/dist/testing/testCommands.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"testCommands.js","sourceRoot":"","sources":["../../src/testing/testCommands.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AACH,OAAO,KAAK,MAAM,OAAO,CAAC;AAC1B,OAAO,EACL,kBAAkB,EAClB,mBAAmB,EACnB,gBAAgB,EAChB,wBAAwB,GACzB,MAAM,wBAAwB,CAAC;AAIhC;;;;;;;;;GASG;AACH,MAAM,UAAU,uBAAuB,CACrC,cAAyC,EACzC,SAGqB;IAErB,cAAc,CAAC,KAAK,GAAG,KAAK,IAAI,EAAE;QAChC,kBAAkB,EAAE,CAAC;IACvB,CAAC,CAAC;IAEF,cAAc,CAAC,UAAU,CAAC,GAAG,KAAK,EAAE,GAAG,IAAc,EAAE,EAAE;QACvD,MAAM,MAAM,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;QACtC,IAAI,CAAC,MAAM,EAAE,CAAC;YACZ,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,MAAM,CAAC,uBAAuB,CAAC,CAAC,CAAC;YACnD,mBAAmB,EAAE,CAAC;YACtB,OAAO;QACT,CAAC;QACD,wBAAwB,CAAC,MAAM,CAAC,CAAC;IACnC,CAAC,CAAC;IAEF,cAAc,CAAC,WAAW,CAAC,GAAG,KAAK,IAAI,EAAE;QACvC,mBAAmB,EAAE,CAAC;IACxB,CAAC,CAAC;IAEF,cAAc,CAAC,UAAU,CAAC,GAAG,KAAK,EAAE,GAAG,IAAc,EAAE,EAAE;QACvD,MAAM,MAAM,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;QACtC,gBAAgB,CAAC,MAAM,IAAI,SAAS,CAAC,CAAC;IACxC,CAAC,CAAC;AACJ,CAAC","sourcesContent":["/**\n * Registers dev-only shell test helpers.\n *\n * Why this file exists:\n * - dev shell commands should stay in one place instead of being spread across the shell setup\n * - test commands should stay in one place instead of being spread across shell setup\n * - the shell should only show explicit maintainer helpers when dev mode is enabled\n */\nimport chalk from \"chalk\";\nimport {\n  runAllTestsCommand,\n  listRegisteredTests,\n  printTestCommand,\n  runRegisteredTestCommand,\n} from \"../commands/TestCmd.js\";\n\ntype CommandFn = (...args: string[]) => Promise<void>;\n\n/**\n * Registers shell test commands.\n * Example: /tests, /test-run <id>.\n *\n * To run tests:\n * - Use `/tests` to run the full test suite.\n * - Use `/test-run <id>` to run a specific test by its ID.\n * - Use `/test-list` to list all available tests.\n * - Use `/test-cmd [id]` to print the command for the full suite or a specific test.\n */\nexport function registerTestingCommands(\n  customCommands: Record<string, CommandFn>,\n  _runQuery: (\n    query: string,\n    options?: { resumeTaskId?: number; continueCurrentTask?: boolean; bindAsCurrentTask?: boolean }\n  ) => Promise<unknown>\n): void {\n  customCommands.tests = async () => {\n    runAllTestsCommand();\n  };\n\n  customCommands[\"test-run\"] = async (...args: string[]) => {\n    const testId = (args[0] ?? \"\").trim();\n    if (!testId) {\n      console.log(chalk.yellow(\"Usage: /test-run <id>\"));\n      listRegisteredTests();\n      return;\n    }\n    runRegisteredTestCommand(testId);\n  };\n\n  customCommands[\"test-list\"] = async () => {\n    listRegisteredTests();\n  };\n\n  customCommands[\"test-cmd\"] = async (...args: string[]) => {\n    const testId = (args[0] ?? \"\").trim();\n    printTestCommand(testId || undefined);\n  };\n}\n"]}